Heterocycle-heterocycle strategies: (2-nitrophenyl)isoxazole precursors to 4-aminoquinolines, 1 H-indoles, and quinolin-4(1 H)-ones

Article abstract of DOI:10.1021/ol400787y

Reductive heterocycle-heterocycle (heterocycle → heterocycle; H-H) transformations that give 4-aminoquinolines, 3-acylindoles, and quinolin-4(1H)-ones from 2-nitrophenyl substituted isoxazoles are reported. When this methodology is applied to 3,5-, 4,5-,

Full text of DOI:10.1021/ol400787y

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
HeterocycleꢀHeterocycle Strategies:
(2-Nitrophenyl)isoxazole Precursors
to 4‑Aminoquinolines, 1H‑Indoles,
and Quinolin-4(1H)‑ones
Keith C. Coffman, Teresa A. Palazzo, Timothy P. Hartley, James C. Fettinger,
Dean J. Tantillo,* and Mark J. Kurth*
Department of Chemistry, University of California, Davis, One Shields Avenue,
Davis, California 95616, United States
mjkurth@ucdavis.edus (synthetic aspects); djtantillo@ucdavis.edu
(theoretical aspects)
Received March 22, 2013
ABSTRACT
Reductive heterocycleꢀheterocycle (heterocycle f heterocycle; HꢀH) transformations that give 4-aminoquinolines, 3-acylindoles, and quinolin-
4(1H)-ones from 2-nitrophenyl substituted isoxazoles are reported. When this methodology is applied to 3,5-, 4,5-, and 3,4-bis(2-
nitrophenyl)isoxazoles, chemoselective heterocyclization gives quinolin-4(1H)-ones, and 4-aminoquinolines, exclusively.
In diversity oriented syntheses,1 heterocycleꢀheterocycle
(HꢀH) strategies, wherein a starting heterocycle is trans-
formed into a new and architecturally different heterocycle,
constitute a powerful means to address diversity in dis-
covery chemistry.2 We postulate that HꢀH strategies can
be particularly useful in providing skeletal diversity and
that this approach uniquely complements the more well
established ‘reagent-based’ and ‘substrate-based’ approaches
to skeletal diversification.3 Indeed, we have recently exploited
an HꢀH strategy in a variety of indazole f indazolone
studies4 and report an extension to isoxazole based systems.
From previous work,5 substituted (3- and 4- nitrophenyl)-
isoxazoles can be selectively reduced to their corresponding
anilines with Zn/HOAc (0 °C, 5 min) and we set out to
explore the possibility of reducing 2-nitrophenylisoxazoles
to other useful heterocycles with Zn/HOAc. Analysis of the
(1) (a) Schreiber, S. L. Science 2000, 287, 1964. (b) Galloway, W. R. J.
D.; Isidro-Llobet, A.; Spring, D. R. Nat. Commun. 2010 DOI: 10.1038/
ncomms1081, 1. (c) Patil, N. T.; Shinde, V. S.; Sridhar, B. Angew. Chem.,
Int. Ed. 2013, 52, 2251. (d) Burke, M. D.; Schreiber, S. L. Angew. Chem.,
Int. Ed. 2004, 43, 46. (e) Tan, D. S. Nat. Chem. Biol. 2005, 1, 74. (f)
Schreiber, S. L. Nature 2009, 457, 153. (g) Thomas, G. L.; Wyatt, E. E.;
Spring, D. R. Curr. Opin. Drug Discovery Dev. 2006, 9, 700.
(2) (a) Guirado, A.; Andreu, R.; Martinez, B.; Bautista, D.; Arellano,
C. R.; Jones, P. G. Tetrahedron 2006, 62, 6172. (b) Fahmy, A. F. M.
ARKIVOC 2006, 7, 395. (c) Guirado, A.; Andreu, R.; Galvez, J.
Tetrahedron Lett. 2003, 44, 3809. (d) Pace, A.; Pibiri, I.; Buscemi, S.;
Vivona, N. Heterocycles 2004, 63, 2627. (e) Piccionello, A. P.; Guarcello,
A.; Buscemi, S.; Vivona, N.; Pace, A. J. Org. Chem. 2010, 75, 8724.
(4) (a) Conrad, W. E.; Rodriguez, K. X.; Nguyen, H. H.; Fettinger,
J. C.; Haddadin, M. J.; Kurth, M. J. Org. Lett. 2012, 14, 3870. (b)
Conrad, W. E.; Fukazawa, R.; Haddadin, M. J.; Kurth, M. J. Org. Lett.
2011, 13, 3138. (c) Donald, M. B.; Conrad, W. E.; Oakdale, J. S.; Butler,
J. D.; Haddadin, M. J.; Kurth, M. J. Org. Lett. 2010, 12, 2524. (d)
Oakdale, J. S.; Solano, D. M.; Fettinger, J. C.; Haddadin, M. J.; Kurth,
M. J. Org. Lett. 2009, 11, 2760.
(5) (a) Coffman, K. C.; Hartley, T. P.; Dallas, J. L.; Kurth, M. J. ACS
Comb. Sci. 2012, 14, 280. (b) Butler, J. D.; Coffman, K. C.; Ziebart,
K. T.; Toney, M. D.; Kurth, M. J. Chem.;Eur. J. 2010, 16, 9002.
ꢀ
ꢀ
(3) (a) Galloway, W. R. J. D.; Diaz-Gavilan, M.; Isidro-Llobet, A.;
Spring, D. R. Angew. Chem., Int. Ed. 2009, 48, 1194. (b) Morton, D.;
Leach, S.; Cordier, C.; Warriner, S.; Nelson, A. Angew. Chem., Int. Ed.
2009, 48, 104. (c) Spandl, R. J.; Bender, A.; Springer, D. R. Org. Biol.
Chem. 2008, 6, 1149. (d) Kumagai, N.; Muncipinto, G.; Schreiber, S. L.
Angew. Chem., Int. Ed. 2006, 45, 3635.
r
10.1021/ol400787y
XXXX American Chemical Society

literature revealed that others had reported related trans-
formations. Specifically, Batra6a et al. reported that a
palladium on carbon-mediated reduction of 3-(2-nitrophenyl)-
isoxazole leads to substituted 4-aminoquinolines (10
examples)6 and Yamanaka et al. reported that Raney
nickel reduction of 4- or 5-(2-nitrophenyl)isoxazoles can
deliver 3-acylindole7 (2 examples) or quinolin-4(1H)-
one7b (1 example) heterocycles, respectively.
As with 2b, all attempts at metal-mediated9 reactions on 3d
[f 5H-indolo[3,2-b]quinolin-11(10H)-one] failed. The re-
ductive heterocyclizations of 1c and 1d proceed by the blue
aniline NH2 attacking the isoxazole-derived imine carbon.
With these calibrating results in hand, wherein both the
carbonyl and imine carbons can serve as the reactive
electrophile, our next objective was to evaluate whether
presumed intermediate 7 (Scheme 2) would undergo
chemoselective heterocyclization of the red aniline NH2
onto the imine or onto the carbonyl carbon. Each hetero-
cyclization of 6 would deliver a different 3-acylindole
when R1 ¼ R2. To address this competition question, both
isomers of 6 (a and b) were prepared by the base-mediated
condensation of the appropriate chlorooxime 4 with the
appropriate 1-(2-nitrophenyl)alkan-2-one 5 (f 6a; f 6b).
In the event, reduction of 6a under Fe/HOAc conditions
gave only indole 8a in 46% yield. Likewise, reduction of 6b
gave only indole 8b in 60% yield (see X-ray in Supporting
Information (SI)).10 Inbothcases, the aniline NH2 cyclized
Scheme 1. HꢀH Chemistry of 3- and 5-(2-Nitrophenyl)-
isoxazoles
This backdrop, coupled with the HꢀH potential of this
strategy, led us to explore this topic further, focusing
initially on using M (= Zn or Fe) in HOAc to trans-
form (2-nitrophenyl)isoxazoles into quinolines/indoles/
quinolin-4(1H)-ones with subsequent studies exploring
the outcomes of reducing bis(2-nitrophenyl)isoxazoles
under these same conditions. We began by preparing
(2-nitrophenyl)isoxazole 1a (Scheme 1) and found that
this substituted isoxazole did not reduce in Zn/HOAc at rt.
Fortunately, we found that Fe powder in neat HOAc at
120 °C resulted in reductive heterocyclization to 4-amino-
quinoline 2a in 77% yield. To extend this methodology,
isoxazole 1a was cleanly converted to 4-bromoisoxazole 1b
(Br2, CCl4, 99% yield) and subsequent Fe/HOAc reduc-
tion delivered 3-bromo-4-aminoquinoline (2b). Attempts
to combine this 1b f 2b HꢀH transformation with sub-
sequent o-haloaniline-based coupling reactions failed.
These reductive heterocyclizations of 1a and 1b proceed
by the green aniline NH2 attacking the isoxazole-derived
carbonyl.
Scheme 2. HꢀH Chemistry of 4-(2-Nitrophenyl)isoxazole
with complete chemoselectivity onto the imine carbon of 7.11
Although intermediate 7 has several possible tautomeric
forms that are all likely accessible under the reaction
conditions12 and may display different modes of reactiv-
ity, the observed chemoselectivity is complete. Although
we are not certain which tautomeric form (or protonated
form thereof) of 7 preferentially reacts in the systems
described herein, we formulate our discussion around
the β-iminocarbonyl 7a.
We next prepared isoxazole 1c by reacting 2-bromo-N-
hydroxybenzimidoyl chloride with 2-nitrophenylacety-
lene.8 Fe/HOAc reduction of isoxazole 1c gave 2-(2-
bromophenyl)-4-quinolin-4(1H)-one 3c in 44% yield. We
found thatBr2 in CCl4 was ineffective atC4bromination of
isoxazole 1c; itwas however readily brominated to 1dusing
NBS (HOAc, cat. H2SO4). Fe/HOAc reduction of 1d gave
3-bromo-2-phenyl-4-quinolin-4(1H)-one 3d (44% yield).
To test the limits of imine vs carbonyl chemoselectivity,
6c was prepared (Scheme 3). Our reasoning was that a
(6) (a) Singh, V.; Yadav, G. P.; Maulik, P. R.; Batra, S. Synthesis
2006, 12, 1995. (b) Casnati, G.; Quilico, A.; Ricca, A.; Finzi, P. V.
Tetrahedron Lett. 1966, 7, 233. (c) Thomson, I.; Torssell, K. B. G. Acta
Chem. Scand., Ser. B 1988, 42, 309. (d) Jensen, S.; Torssell, K. B. G. Act.
Chem. Scand. 1995, 49, 53.
(7) (a) Uchiyama, D.; Yabe, M.; Kameyama, H.; Sakamoto, T.;
Kondo, Y.; Yamanaka, H. Heterocycles 1996, 43, 1301. (b) Sakamoto,
T.; Kondo, Y.; Uchiyama, D.; Yamanaka, H. Tetrahedron 1991, 47,
5111.
(8) (a) Guggenheim, K. G.; Butler, J. D.; Painter, P. P.; Lorsbach,
B. A.; Tantillo, D. J.; Kurth, M. J. J. Org. Chem. 2011, 76, 5803.
(b) Meng, L.; Lorsbach, B. A.; Sparks, T. C.; Fettinger, J. C.; Kurth,
M. J. J. Comb. Chem. 2010, 12, 129.
(9) For a review of Buchwald chemistry, see: Fischer, C.; Koenig, B.
Beilstein J. Org. Chem. 2011, 7, 59.
(10) CCDC 926644 (8b), 926905 (10a), and 926645 (10b) contain the
supplementary crystallographic data for this paper. These data can be
obtained free of charge from the Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
(11) See Supporting Information for explanation of chemoselectivity.
(12) Tautomer energies of 7a, 7b, and 7c (R3 = Ph, R4 = H, R5 = Ph)
were computed (B3LYP/6-31þG(d,p); see SI for complete details).
It was determined that 7c had the lowest overall energy, while 7a and 7b
were 10.1 and 5.2 kcal/mol higher, respectively.
B
Org. Lett., Vol. XX, No. XX, XXXX

2-methoxyphenyl moiety (R2 in 7; Scheme 2) would donate
electron density into the imine of 7 and reduce its electro-
philicity. Treating 6c with Fe/HOAc at 120 °C resulted in
the formation of two indole products, 10a and 10b in a
combined yield of 96% and in a 60:40 ratio, respectively.
Thus, even in this “reactivity-skewed” case, the aniline
NH2 prefers to attack the imine (9a f 10a), in contrast to
Yamanaka’s one asymmetrical example.7a The structures
of 10a and 10b were established by X-ray crystallography
(see SI).10
effects of different conjugation pathways on the acidity/
basicity (and, potentially, iron-binding ability) of key
functional groups, and different tether lengths between
potential nucleophiles and electrophiles (6-exo-trig vs
5-exo-trig).15 Despite these complications, the HꢀH re-
ductive heterocyclizations depicted in Schemes 2 and 4
each proceed to give only one new heterocyclic product.
Scheme 4. HꢀH Chemistry of 3,5-Bis(2-nitrophenyl)isoxazole
Scheme 3. HꢀH Chemistry of 4-(2-Nitrophenyl)isoxazole
With these Fe/HOAc-mediated (2-nitrophenyl)isoxaz-
ole f 4-aminoquinoline/indole/quinolin-4(1H)-one HꢀH
results in hand, we next explored the reductive heterocycli-
zation chemistry of 3,5-, 3,4-, and 4,5-bis(2-nitrophenyl)-
isoxazoles as it appeared these systems would confront new,
as yet unexplored chemoselectivity questions. For example,
reductive heterocyclization of 3,5-bis-(2-nitrophenyl)-
isoxazole 11 (Scheme 4) could produce quinolin-4(1H)-
one (w heterocyclization to 12 involving the imine via 12a)
and/or 4-aminoquinoline (w heterocyclization to 13 in-
volving the carbonyl via 13a) products. To probe this
question, 11 was synthesized.
The complete imine vs carbonyl chemoselectivity ob-
served for 11 suggested that 3,4-bis(2-nitrophenyl)-
isoxazole 14 (Scheme 5) should produce 3-acyl-1H-indole
17 (and, perhaps, its condensation analogue 18). To test
this hypothesis, 1-(2-nitrophenyl)-propan-2-one was
condensed with N-hydroxy-2-nitrobenzimidoyl chloride
(NaH in dry THF). Reductive heterocyclization of 14
with Fe/HOAc gave none of the anticipated imine-derived
indole 17; rather, 4-aminoquinoline 15 and its acylated
analogue16(69% combinedyield; 0.2:1 ratio, respectively)
were produced.
Reductive heterocyclization of 11 with Fe/HOAc
gave quinolin-4(1H)-one 12 (22% isolated þ several uni-
dentified spots in the crude TLC); we suspected that
in situ formation of acetic anhydride and subsequent
N-acylation(s) of the reaction products complicated this
reaction mixture.13 It was found that changing the acid
source from HOAc to aq. NH4Cl increased the yield of 12
to 56%. Thus, 11, upon reduction, leads to chemoselective
attack by the blue aniline NH2 onto the imine carbon
(12a f 12) to the complete exclusion of the green aniline
NH2 onto the carbonyl carbon (13a f 13). Importantly,
no 4-aminoquinoline 13 was detected in either reaction.14
Again, understanding the heterocyclization reactivity
of bis(2-nitrophenyl)isoxazole derived intermediates is
complicated by the viability of various β-iminocarbonyl
tautomers, their numerous conformations and configura-
tions, multiple possible hydrogen bonding arrays, the
Obtaining 16 supports the suspected in situ formation
of acetic anhydride under these conditions.13 Since neither
indole 17 or 18 was obtained, we believe that this hetero-
cyclization is prevented because the imine in 17a (f 17)
suffers deactivation by conjugation with the green aniline
NH2 (similar to 6c; Scheme 3). Accepting that as an
explanation for no formation of 17, why was 4-aminoqui-
noline 15 formed to the complete exclusion of indole 19?
We hypothesize that the red aniline NH2 (derived from the
red nitro moiety) is the most basic site in the bis-reduction
intermediate derived from 14: this nitrogen’s lone pair is
not delocalized into the β-iminocarbonyl system whereas
the green aniline NH2 would be. Consequently, in acetic
acid at 120 °C, the red aniline is protonated (see 15a in
Scheme 5) and not available for the required nucleophilic
attack. Indeed, this factor may also explain the lack of
formation of 17.
These insights with 14 led us to predict that reductive
heterocyclization of the final system in this series, i.e.,
4,5-bis-(2-nitrophenyl)isoxazole 20 (Scheme 6), should
produce quinolin-4(1H)-one 21a (w reductive heterocycli-
zation of 20 to 21a would involve the imine carbon and the
blue conjugated aniline NH2) via intermediate 20a. To test
(13) For related studies, see: (a) Yamada, Y.; Segawa, M.; Sato, F.;
Kojima, T.; Sato, S. J. Mol. Catal. A: Chem. 2011, 346, 79. (b) Karimi,
E.; Teixeira, I. F.; Ribeiro, L. P.; Gomez, A.; Lago, R. M.; Penner, G.;
Kycia, S. W.; Schlaf, M. Catal. Today 2012, 190, 73. (c) Squibb, E. R.
J. Am. Chem. Soc. 1895, 17, 187 and references cited therein.
(14) HRMS calculated for 12 [C15H12N2O þ H]þ, 237.1022; found,
237.1030. HRMS results along with NMR spectra confirmed the
identity of 12. Fortunately, 13 has one less proton ([M þ H]þ:
236.1182) leading to the conclusion that one product was formed
exclusively.
(15) The results of quantum chemical calculations on model systems
can be found in the SI.
Org. Lett., Vol. XX, No. XX, XXXX
C

Scheme 5. HꢀH Chemistry of 3,4-Bis(2-nitrophenyl)isoxazole
Scheme 6. HꢀH Chemistry of 4,5-Bis(2-nitrophenyl)isoxazole
In this work, we have shown that Fe/HOAc reduction
of (2-nitrophenyl)- and bis(2-nitrophenyl)isoxazoles leads
with great chemoselectivity to a variety of useful hetero-
cycles via heterocycleꢀheterocycle (HꢀH) transforma-
tions. Although the reactivities of the systems described
above are influenced by many factors, often juxtaposed in
effect, two empirical guidelines for predicting the products
in these systems have emerged: (1) chemoselective hetero-
cyclization onto the imine carbon is preferred unless it is
deactivated;(2) insystemswithcompeting anilinemoieties,
the less basic NH2 is the favored nucleophile. These guide-
lines will, no doubt, prove useful in facilitating the design
of additional HꢀH synthetic reactions of this type.
this prediction, 20 was synthesized by condensation of
N-hydroxyacetimidoyl chloride onto 1,2-bis(2-nitro-
phenyl)ethanone (NaH in dry THF). Fe/HOAc reduction
of 20 proceeded to give quinolin-4(1H)-ones 21a and 21b
(87% total yield) in a 1:1.1 ratio, respectively, confirming
formation of acetic anhydride in situ.13 Importantly, in-
doles 22 and 23 were not obtained as their formation
is precluded by formation of the protonated red aniline
NH2 in 20a.
Since the assignment of structure 21a was complicated
bythe expectation that 22and 23would have similar NMR
spectra and the same mass, chemical shift calculations
(using the multistandard approach)16 were performed
using DFT17 (see SI for details). Mean absolute deviations
(MAD) of the chemical shifts relative to experimental data
were 0.2, 0.4, and 0.6 ppm (21a, 22, and 23 respectively) for
1H NMR. The MAD for the 13C NMR were 2.8, 4.8, and
4.8 ppm, respectively.18 The MAD for structure 21a are the
lowest and within the range typically found for correctly
assigned structures,16 suggesting that 21a is indeed the
product of the reduction.19
Acknowledgment. We thank the Tara K. Telford Fund
for Cystic Fibrosis Research at UC Davis, the National
Institutes of Health (GM0891583, DK072517, and
RR1973), and the National Science Foundation [CHE-
0443516/0449845/9808183, and DBIO-722538 for NMR
spectrometers, CHE-0840444 for the Dual source X-ray
diffractometer, CHE-030089 for computer support] for
generous support. We also thank Kelli M. Gottlieb
(University of California, Davis) for collection of HRMS
data and Mike Lodewyk (University of California, Davis)
for his advice with the calculations.
(16) Lodewyk, M.; Siebert, M. R.; Tantillo, D. J. Chem. Rev. 2012,
112, 1839.
(17) Frisch, M. J. Gaussian 09 (revision B.01): See SI for full author list.
(18) Figures are in the SI.
(19) Typical 3-acylindole 13C CdO shifts are ∼190 ppm, while
quinolin-4-ones are ∼175 ppm. The carbonyl shift from the reduction
of 4,5-bis(2-nitrophenyl)isoxazole was 175.0 giving further evidence that
quinolin-4-one 21a was formed in the reduction. For indole related
examples, see: (a) Stokes, B. J.; Liu, S.; Driver, T. G. J. Am. Chem. Soc.
2011, 113, 4702. (b) Guchhait, S. K.; Kashyap, M.; Kamble, H. J. Org.
Chem. 2011, 76, 4853. For quinolin-4-one related examples, see: (c)
Cross, R. M.; Manetsch, R. J. Org. Chem. 2010, 75, 8654. (d) Luo, F. T.;
Ravi, V. K.; Xue, C. Tetrahedron 2006, 62, 9365.
Supporting Information Available. Experimental pro-
cedures, full spectroscopic data for all new compounds,
and X-ray crystallographic data for 8b, 10a, and 10b. This
material is available free of charge via the Internet at
http://pubs.acs.org.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX